DepartmentofComputerScience

Title: DepartmentofComputerScience

Research Question: How can we formalize and reason about anonymity in multi-agent systems?

Methodology: The authors use the modal logic of knowledge within the context of the runs-and-systems framework, similar to their earlier work on secrecy. They define several types of anonymity with respect to agents, actions, and observers in multi-agent systems, and relate these definitions to other concepts of information hiding. They also provide probabilistic definitions of anonymity to quantify an observer's uncertainty about the system's state.

Results: The authors present several definitions of anonymity, including the concept of indistinguishability, where an observer cannot distinguish between the actions of different agents, and the concept of unobservability, where an observer cannot determine the actions performed by an agent. They relate these definitions to other concepts of information hiding and provide probabilistic definitions to measure anonymity quantitatively.

Implications: The authors' framework provides a clear and formal way to reason about anonymity in multi-agent systems. This can help in the design and analysis of systems that aim to protect user privacy and maintain information security. The probabilistic definitions of anonymity can also be useful in quantifying the level of privacy provided by a system.

Link to Article: https://arxiv.org/abs/0402042v1 Authors: arXiv ID: 0402042v1